[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Update on comet collision vs. KT impact theories

As promised, here is the scoop on the Shoemaker-Levy 9 impact from
the International Astronomical Union general assembly in The Hague
the last couple of weeks. I sneaked away from galaxy sessions for
one of the Jupiter/SL9 sessions. Much of what follows is taken from talks
by Mike A'Hearn (University of Maryland), Heidi Hammel (MIT), John
Clarke (University of Michigan), and Keith Noll (Space Telescope Science
Institute). I have added a bit of material on the comet orbital
dynamics and which directions might be needed to scale these events in
understanding impacts on Earth (though this does go beyond my specialty
so I have exercised some restraint).

The comet's orbit : A'Hearn pointed out that we were properly discussing
the former Jovian satellite Shoemaker-Levy 9. While prediscovery
observations after the tidal breakup have been found, none has
turned up before then. This raises the possibility that there could
be more such objects currently in Jovian (Jovicentric?) orbits but
too faint to see. The latest orbit passed in July 1992 within 1.5
Jupiter radii, carrying the comet inside the classical Roche limit for
tidal breakup for 1.5 hours (Gravity - not just a good idea. It's the law),
and out to apojove (our word for the month) 0.3 astronomical units
(45 million km) from the planet. At this distance, solar perturbations
are very important (which is what made the orbit unstable). For
such a highly inclined, retrograde orbit in that direction, solar
perturbations decreased the orbital angular momentum, with the now
well-known spectacular results. The comet was in Jupiter orbit for several
decades; available observations cannot be more specific because the
previous orbit determinations are very sensitive to the perijove
at each encounter, which in turn is very sensitive to small errors in
the current orbit (in fact, it has n discussed as satisfying the
criteria for chaos over this period).
    This temporary capture works because Jupiter (a) has numerous
satellites, some massive, and more important (b) sits in the solar
gravitational well. The parameter space for Earth to capture such an
object is much smaller. In fact, Clyde Tombaugh made an extensive search
for small natural Earth satellites shortly before such a search was
rendered impossible by artificial satellites (obviously with null results).
However, Earth can certainly disrupt comets by tidal forces - the moon has
crater chains that look much like SL9 would have produced. A good
example is Davy Catena (formerly the Davy Y chain) near the center of the
lunar earthside. Good images can be found on page 97 of NASA SP-200, The Moon 
as Viewed by Lunar Orbiter, and plate 297 (upper right corner) of
NASA SP-206, Lunar Orbiter Photographic Atlas of the Moon. 
    In this case, the impact speed was dominated by Jupiter's escape
velocity, since the object was gravitationally bound to Jupiter (but
just barely). This is not likely to be the case for Earth impacts, as
has been pointed out by others. Jupiter's escape velocity is larger than
circular-orbit velocity at its distance from the Sun, while the opposite
holds in our case. Comets are likely to have higher impact speeds than
asteroids, which are in more nearly circular and direct (not retrograde)

Was this a comet or an asteroid before breakup? Not a trivial question.
It certainly acted cometary after breakup - dust tails, disappearing and
brightening/fading fragments. On the other hand, we don't know that 
asteroids don't act this way when broken up, having never seen such a
thing happen. When sideswiped, asteroids do produce dust trails (seen
in the far-infrared by their own thermal emission). SL9 showed (as far
as I know) no evidence that any gas was being released from the fragments,
and for that matter no dust except what had been released at breakup.
New observations from the Kuiper Airborne Observatory showed water in some
of the impact plumes, and the time behavior has been taken to suggest that
it came from the fragments. If so, this sounds like a genuine comet.
     SL9 was certainly cometary when it hit Jupiter. Some very impressive
(and difficult) observations of fragments only a few hours before
impact show the dust clouds stretching out along the radius vector,
strongest toward Jupiter, as expected from tidal acceleration (and
the string of pearls stretched some millions of km at this point for
the same reason).

Mass of the impacting fragments: from the energy budget of the impact
fireballs, the largest ones had masses of order 10^14 g. For solid
ice, this gives a diameter of about  0.6 km, larger if it had a
porous or "snowball" structure. A nickel/iron meteor with the
same punch would have a diameter smaller by the cube root of the
density ratio, or about 1.6x. The energy released in the larger impacts
was about 2x10^27 ergs (or for all you ex-cold warriors out there, I make it
close to 10^5 megatons of TNT). This scales as the square of the impact 
velocity, so a similar impactor would (likely but not guaranteed)
give a smaller energy release on Earth.
     One real puzzle is the lack of obvious meteor phenomena. Comparison
of times of flashes seen by Galileo with Earth-based IR observations
shows that no brilliant flash occurred on atmospheric entry until
the fragments dropped below the cloudtops - stealth comets! The only
possible exceptions were, curiously, the fragments that left little or
no trace of the cloudtops, and were (even more curiously) offset from
the main line of fragments (density effects???) - some of these have
unconfirmed reports of flashed reflected from the large Jovian

Where did they explode? The upper Jovian atmosphere has three major cloud
layers in a hydrogen-rich matrix. Lacking a sea level, it is usual to
reference height to the one-bar pressure level. On this scale,
the visible cloud tops are ammonia at about +50 km, NH_3SH at about 0 km,
and water clouds at about -20 km. No explosions seemed to dredge up
significant water, and indirect indications (such as the center of
shock waves compared to entry points seen in HST images) agree that
the fragments exploded perhaps a couple of hundred km below the top clouds
but (from spectroscopy) above the water clouds (a small amount of
H2O did eventually turn up in spectra of the impact sites). In an
Earth impact, the fragments would have hit the surface instead of
such an air burst, as the column density and pressure are less through
our entire atmosphere than the fragments traversed.
     The explosions channeled much of the rising material back along the
low-density channel left by the incoming fragment (predicted by at
least one simulation at Sandia Labs - in Shock Waves 4, 47., 1994 by
Crawford et al.). The expanding fireballs rose to heights exceeding
3000 km above the cloudtops and were briefly seen in visible light 
by their own thermal emission (directly showing temperatures of several
thousand K). Spectra of these sites showed that even refractory elements
such as sodium, magnesium, iron, and silicon were vaporized (and if I
interpret my notes properly dissociated into atomic form  rather than
remaining bound into molecules).
     The dark eject are still visible (at least the major impacts). The
material (still unclear whether of cometary or Jovian origin) has
a vertical extent of hundreds of km, known from (1) imaging through
filters tuned on and off a strong methane absorption wavelength,
which serves as a convenient depth gauge, and (2) obvious parallax
against underlying cloud features as Jupiter rotates. The mass of
material (assuming small particles) is comparable to the estimated
impact mass - I saw no good arguments that this is more than coincidence,
though. It was sobering to see quite directly that each of these impacts
could have more than blanketed the Earth with absorbing material
(specially with our lower surface gravity), since the bigger impact 
spots spread over a couple of times Earth's surface area. The transport
of the material would be more complicated here, with the planet's
curvature being a stronger factor. Broad energy considerations (best I
can tell) suggest that little material would reach escape velocity
even in a terrestrial impact. The column density of absorbing material
is low by terrestrial standards (estimated at 5x10^18 atoms/cm^2)
but the integrated absorption of sunlight is rather efficient
(as we see, since the dark spots are so prominent; but recall that we
see double the line-of-sight absorption, since the sunlight we see
by and large passes through the absorbing material once on the way
down and once again on the way up).
     The rapid winds above the cloudtops (never before directly measured,
by the way) are disrupting the absorbing clouds, but slowly - at this
rate the bigger ones will last for months. Unless the different 
atmospheric structures do something I don't see, this is pretty direct
evidence that impacts would produce an absorbing shround around the
Earth of duration at least a year (not even counting proposed 
constituents like soot that don't enter on Jupiter).

There were also interesting effects on the Jovian magnetic field,
though they don't enter for the present purposes  archosaurs
migrated following magnetic field lines (harrumph). New auroral
zones were observed fed by particles released at several impacts,
and the radio emission from Jupiter's radiation belts was pumped
up during the string of impacts.

The upshot is that, for understanding the effects of any impacts
in terrestrial history, the models for energy release and
atmospheric blanketing don;t seem to have too wide of the mark.
Now - is there an atmospheric physicist out there who can tell us more
about how these phenomena would scale to the terrestrial atmosphere?

Bill Keel                           Astronomy, University of Alabama